Significant variations in the twinning rate can be observed both in the long term and in the short term. Attempts to explain such variation have been based variously on coital frequency, declining sperm counts, a fertility advantage in twin-prone women, variation in early loss of embryos, and other phenomena. Evidence is being reviewed, including recent vital statistics and current views of the physiology of reproduction, in the expectation that a decision can be made as to the probable explanation of twinning rate variations.